syntaxtree.py 41.2 KB
Newer Older
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
# syntaxtree.py - syntax tree classes for DHParser
#
# Copyright 2016  by Eckhart Arnold (arnold@badw.de)
#                 Bavarian Academy of Sciences an Humanities (badw.de)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
# implied.  See the License for the specific language governing
# permissions and limitations under the License.
17
18
19


"""
20
21
Module ``syntaxtree`` defines the ``Node``-class for syntax trees as well
as an abstract base class for parser-objects. The latter is defined
eckhart's avatar
eckhart committed
22
here, because node-objects refer to parser-objects. All concrete
23
24
25
parser classes are defined in the ``parse`` module.
"""

26

Eckhart Arnold's avatar
Eckhart Arnold committed
27
import collections.abc
28
from collections import OrderedDict
eckhart's avatar
eckhart committed
29
import copy
30

eckhart's avatar
eckhart committed
31
from DHParser.error import Error, ErrorCode, linebreaks, line_col
32
from DHParser.stringview import StringView
eckhart's avatar
eckhart committed
33
from DHParser.toolkit import re, typing
eckhart's avatar
eckhart committed
34
from typing import Callable, cast, Iterator, List, AbstractSet, Set, Dict, Union, Tuple, Optional
di68kap's avatar
di68kap committed
35

36

37
38
39
40
41
42
__all__ = ('ParserBase',
           'WHITESPACE_PTYPE',
           'TOKEN_PTYPE',
           'MockParser',
           'ZombieParser',
           'ZOMBIE_PARSER',
eckhart's avatar
eckhart committed
43
           'ZOMBIE_NODE',
44
           'Node',
di68kap's avatar
di68kap committed
45
           'RootNode',
eckhart's avatar
eckhart committed
46
           'ZOMBIE_ROOTNODE',
47
           'parse_sxpr',
48
49
50
           'parse_xml',
           'flatten_sxpr',
           'flatten_xml')
Eckhart Arnold's avatar
Eckhart Arnold committed
51
52


53
54
55
56
57
58
59
60
61
62
63
64
65
#######################################################################
#
# parser base and mock parsers
#
#######################################################################


class ParserBase:
    """
    ParserBase is the base class for all real and mock parser classes.
    It is defined here, because Node objects require a parser object
    for instantiation.
    """
66

67
    __slots__ = 'name', 'ptype'
Eckhart Arnold's avatar
Eckhart Arnold committed
68

69
70
71
    def __init__(self,):  # , pbases=frozenset()):
        self.name = ''  # type: str
        self.ptype = ':' + self.__class__.__name__  # type: str
72
73

    def __repr__(self):
eckhart's avatar
eckhart committed
74
        return self.name + self.ptype
75
76
77
78

    def __str__(self):
        return self.name + (' = ' if self.name else '') + repr(self)

79
80
81
    def __call__(self, text: StringView) -> Tuple[Optional['Node'], StringView]:
        return None, text

82
83
    @property
    def repr(self) -> str:
di68kap's avatar
di68kap committed
84
        """Returns the parser's name if it has a name and repr()"""
85
86
        return self.name if self.name else repr(self)

87
    def reset(self):
88
        """Resets any parser variables. (Should be overridden.)"""
89
90
        pass

eckhart's avatar
eckhart committed
91
    def grammar(self) -> Optional[object]:
92
93
        """Returns the Grammar object to which the parser belongs. If not
        yet connected to any Grammar object, None is returned."""
94
95
96
        return None

    def apply(self, func: Callable) -> bool:
eckhart's avatar
eckhart committed
97
        """Applies the function `func` to the parser. Returns False, if
98
        - for whatever reason - the functions has not been applied, True
eckhart's avatar
eckhart committed
99
        otherwise."""
100
101
        return False

102
103

WHITESPACE_PTYPE = ':Whitespace'
104
TOKEN_PTYPE = ':Token'
105
106
107
108
109
110


class MockParser(ParserBase):
    """
    MockParser objects can be used to reconstruct syntax trees from a
    serialized form like S-expressions or XML. Mock objects can mimic
eckhart's avatar
eckhart committed
111
    different parser types by assigning them a `ptype` on initialization.
112
113
114
115
116

    Mock objects should not be used for anything other than
    syntax tree (re-)construction. In all other cases where a parser
    object substitute is needed, chose the singleton ZOMBIE_PARSER.
    """
117

Eckhart Arnold's avatar
Eckhart Arnold committed
118
119
    __slots__ = ()

120
121
    def __init__(self, name='', ptype=''):  # , pbases=frozenset()):
        assert not ptype or ptype[0] == ':'
122
123
        super().__init__()
        self.name = name
di68kap's avatar
di68kap committed
124
125
        if ptype:
            self.ptype = ptype  # or ':' + self.__class__.__name__
126
127
128
129
130
131
132
133
134
135
136
137


class ZombieParser(MockParser):
    """
    Serves as a substitute for a Parser instance.

    ``ZombieParser`` is the class of the singelton object
    ``ZOMBIE_PARSER``. The  ``ZOMBIE_PARSER`` has a name and can be
    called, but it never matches. It serves as a substitute where only
    these (or one of these properties) is needed, but no real Parser-
    object is instantiated.
    """
138

139
    alive = False
Eckhart Arnold's avatar
Eckhart Arnold committed
140
    __slots__ = ()
141
142

    def __init__(self):
143
        super(ZombieParser, self).__init__()
144
145
        assert not self.__class__.alive, "There can be only one!"
        assert self.__class__ == ZombieParser, "No derivatives, please!"
146
        self.name = "__ZOMBIE__"
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
        self.__class__.alive = True

    def __copy__(self):
        return self

    def __deepcopy__(self, memo):
        return self

    def __call__(self, text):
        """Better call Saul ;-)"""
        return None, text


ZOMBIE_PARSER = ZombieParser()


#######################################################################
#
# syntaxtree nodes
#
#######################################################################


170
ChildrenType = Tuple['Node', ...]
Eckhart Arnold's avatar
Eckhart Arnold committed
171
NoChildren = cast(ChildrenType, ())  # type: ChildrenType
eckhart's avatar
eckhart committed
172
StrictResultType = Union[ChildrenType, StringView, str]
173
ResultType = Union[ChildrenType, 'Node', StringView, str, None]
174
175


Eckhart Arnold's avatar
Eckhart Arnold committed
176
def flatten_sxpr(sxpr: str) -> str:
177
178
    """
    Returns S-expression ``sxpr`` as a one-liner without unnecessary
Eckhart Arnold's avatar
Eckhart Arnold committed
179
180
181
    whitespace.

    Example:
Eckhart Arnold's avatar
Eckhart Arnold committed
182
    >>> flatten_sxpr('(a\\n    (b\\n        c\\n    )\\n)\\n')
Eckhart Arnold's avatar
Eckhart Arnold committed
183
184
    '(a (b c))'
    """
185

186
    return re.sub(r'\s(?=\))', '', re.sub(r'\s+', ' ', sxpr)).strip()
Eckhart Arnold's avatar
Eckhart Arnold committed
187
188


189
def flatten_xml(xml: str) -> str:
190
191
    """
    Returns an XML-tree as a one liner without unnecessary whitespace,
192
    i.e. only whitespace within leaf-nodes is preserved.
193
194
    A more precise alternative to `flatten_xml` is to use Node.as_xml()
    ans passing a set containing the top level tag to parameter `inline_tags`.
195
    """
196

197
198
199
200
    # works only with regex
    # return re.sub(r'\s+(?=<\w)', '', re.sub(r'(?<=</\w+>)\s+', '', xml))
    def tag_only(m):
        return m.groupdict()['closing_tag']
201
    return re.sub(r'\s+(?=<[\w:])', '', re.sub(r'(?P<closing_tag></:?\w+>)\s+', tag_only, xml))
202
203


eckhart's avatar
eckhart committed
204
RX_AMP = re.compile(r'&(?!\w+;)')
205
206


Eckhart Arnold's avatar
Eckhart Arnold committed
207
class Node(collections.abc.Sized):
208
209
210
    """
    Represents a node in the concrete or abstract syntax tree.

211
212
    TODO: Add some documentation and doc-tests here...

213
    Attributes:
214
215
        tag_name (str):  The name of the node, which is either its
            parser's name or, if that is empty, the parser's class name
216

217
218
219
        result (str or tuple):  The result of the parser which
            generated this node, which can be either a string or a
            tuple of child nodes.
220

221
222
        children (tuple):  The tuple of child nodes or an empty tuple
            if there are no child nodes. READ ONLY!
223

eckhart's avatar
eckhart committed
224
225
226
227
        content (str):  Yields the contents of the tree as string. The
            difference to ``str(node)`` is that ``node.content`` does
            not add the error messages to the returned string.

Eckhart Arnold's avatar
Eckhart Arnold committed
228
        parser (Parser):  The parser which generated this node.
229
230
            WARNING: In case you use mock syntax trees for testing or
            parser replacement during the AST-transformation: DO NOT
Eckhart Arnold's avatar
Eckhart Arnold committed
231
232
            rely on this being a real parser object in any phase after
            parsing (i.e. AST-transformation and compiling), for
233
            example by calling ``isinstance(node.parer, ...)``.
234

235
236
237
        len (int):  The full length of the node's string result if the
            node is a leaf node or, otherwise, the concatenated string
            result's of its descendants. The figure always represents
238
            the length before AST-transformation and will never change
239
            through AST-transformation. READ ONLY!
240

241
242
        pos (int):  the position of the node within the parsed text.

Eckhart Arnold's avatar
Eckhart Arnold committed
243
            The value of ``pos`` is -1 meaning invalid by default.
244
            Setting this value will set the positions of all child
Eckhart Arnold's avatar
Eckhart Arnold committed
245
            nodes relative to this value.
246
247

            To set the pos values of all nodes in a syntax tree, the
Eckhart Arnold's avatar
Eckhart Arnold committed
248
            pos value of the root node should be set to 0 right
249
250
            after parsing.

Eckhart Arnold's avatar
Eckhart Arnold committed
251
            Other than that, this value should be considered READ ONLY.
252
253
            At any rate, it should only be reassigned during the parsing
            stage and never during or after the AST-transformation.
254

255
256
        errors (list):  A list of all errors that occured on this node.

257
258
        attr (dict): An optional dictionary of XML-attr. This
            dictionary is created lazily upon first usage. The attr
259
260
            will only be shown in the XML-Representation, not in the
            S-Expression-output.
261
    """
262

Eckhart Arnold's avatar
Eckhart Arnold committed
263
    __slots__ = '_result', 'children', '_len', '_pos', 'parser', 'errors', '_xml_attr', '_content'
264

265
266
267
    def __init__(self, parser, result: ResultType, leafhint: bool = False) -> None:
        """
        Initializes the ``Node``-object with the ``Parser``-Instance
268
269
        that generated the node and the parser's result.
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
270
        self.errors = []                # type: List[Error]
eckhart's avatar
eckhart committed
271
        self._pos = -1                  # type: int
272
        # Assignment to self.result initializes the attr _result, children and _len
di68kap's avatar
di68kap committed
273
        # The following if-clause is merely an optimization, i.e. a fast-path for leaf-Nodes
274
        if leafhint:
eckhart's avatar
eckhart committed
275
            self._result = result       # type: StrictResultType  # cast(StrictResultType, result)
eckhart's avatar
eckhart committed
276
            self._content = None        # type: Optional[str]
277
            self.children = NoChildren  # type: ChildrenType
278
            self._len = -1              # type: int  # lazy evaluation
279
280
        else:
            self.result = result
281
282
        self.parser = parser or ZOMBIE_PARSER

eckhart's avatar
eckhart committed
283
284
285
286
287
288
289
290
291
292
293
    def __deepcopy__(self, memo):
        if self.children:
            duplicate = self.__class__(self.parser, copy.deepcopy(self.children), False)
        else:
            duplicate = self.__class__(self.parser, self.result, True)
        duplicate.errors = copy.deepcopy(self.errors) if self.errors else []
        duplicate._pos = self._pos
        duplicate._len = self._len
        if hasattr(self, '_xml_attr'):
            duplicate._xml_attr = copy.deepcopy(self._xml_attr)
        return duplicate
Eckhart Arnold's avatar
Eckhart Arnold committed
294

295
    def __str__(self):
296
        s = "".join(str(child) for child in self.children) if self.children else self.content
eckhart's avatar
eckhart committed
297
        if self.errors:
298
            return ' <<< Error on "%s" | %s >>> ' % \
eckhart's avatar
eckhart committed
299
                   (s, '; '.join(e.message for e in self.errors))
300
        return s
301

Eckhart Arnold's avatar
Eckhart Arnold committed
302

303
304
305
306
    def __repr__(self):
        mpargs = {'name': self.parser.name, 'ptype': self.parser.ptype}
        parg = "MockParser({name}, {ptype})".format(**mpargs)
        rarg = str(self) if not self.children else \
eckhart's avatar
eckhart committed
307
            "(" + ", ".join(repr(child) for child in self.children) + ")"
308
        return "Node(%s, %s)" % (parg, rarg)
309

Eckhart Arnold's avatar
Eckhart Arnold committed
310

311
    def __len__(self):
312
        if self._len < 0:
di68kap's avatar
di68kap committed
313
            self._len = sum(len(child) for child in self.children) \
314
                if self.children else len(self._result)
315
316
317
318
319
320
321
322
        return self._len


    def __bool__(self):
        # A node that is not None is always True, even if it's empty
        return True


323
    def __eq__(self, other):
324
325
326
        """
        Equality of nodes: Two nodes are considered as equal, if their tag
        name is the same and if their results are equal.
327
328
329
330

        Note: It is not required that two nodes have the same errors attached.
        In case you need to check for error equality as well, compare a
        serialization that includes error messages, as_sxpr() will do!
331
        """
332
        return self.tag_name == other.tag_name and self.result == other.result
333

Eckhart Arnold's avatar
Eckhart Arnold committed
334

335
    def __hash__(self):
336
        return hash(self.tag_name)
337

Eckhart Arnold's avatar
Eckhart Arnold committed
338

339
340
341
    def __getitem__(self, index_or_tagname: Union[int, str]) -> Union['Node', Iterator['Node']]:
        """
        Returns the child node with the given index if ``index_or_tagname`` is
342
        an integer or the first child node with the given tag name. Examples::
343

344
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
345
346
            >>> flatten_sxpr(tree[0].as_sxpr())
            '(b "X")'
347
348
            >>> flatten_sxpr(tree["X"].as_sxpr())
            '(X (c "d"))'
349
350
351
352

        Args:
            index_or_tagname(str): Either an index of a child node or a
                tag name.
353
        Returns:
354
355
            Node: All nodes which have a given tag name.
        """
356
357
358
        if self.children:
            if isinstance(index_or_tagname, int):
                return self.children[index_or_tagname]
359
            else:
360
361
362
363
364
                for child in self.children:
                    if child.tag_name == index_or_tagname:
                        return child
                raise KeyError(index_or_tagname)
        raise ValueError('Leave nodes have no children that can be indexed!')
365
366
367
368


    def __contains__(self, tag_name: str) -> bool:
        """
369
        Returns true if a child with the given tag name exists.
370
        Args:
371
372
            tag_name (str): tag_name which will be searched among to immediate
                descendants of this node.
373
374
375
376
        Returns:
            bool:  True, if at least one descendant node with the given tag
                name exists, False otherwise
        """
377
378
379
380
381
        # assert isinstance(tag_name, str)
        if self.children:
            for child in self.children:
                if child.tag_name == tag_name:
                    return True
382
            return False
383
        raise ValueError('Leave node cannot contain other nodes')
384
385


386
387
388
389
390
391
392
393
394
395
396
397
398
399
    def get(self, index_or_tagname: Union[int, str],
            surrogate: Union['Node', Iterator['Node']]) -> Union['Node', Iterator['Node']]:
        """Returns the child node with the given index if ``index_or_tagname``
        is an integer or the first child node with the given tag name. If no
        child with the given index or tag_name exists, the ``surrogate`` is
        returned instead. This mimics the behaviour of Python's dictionary's
        get-method.
        """
        try:
            return self[index_or_tagname]
        except KeyError:
            return surrogate


400
    @property   # this needs to be a (dynamic) property, in case sef.parser gets updated
401
    def tag_name(self) -> str:
402
403
404
405
406
407
        """
        Returns the tage name of Node, i.e. the name for XML or
        S-expression representation. By default the tag name is the
        name of the node's parser or, if the node's parser is unnamed, the
        node's parser's `ptype`.
        """
408
        return self.parser.name or self.parser.ptype
409

Eckhart Arnold's avatar
Eckhart Arnold committed
410

411
    @property
Eckhart Arnold's avatar
Eckhart Arnold committed
412
    def result(self) -> StrictResultType:
413
414
415
416
417
        """
        Returns the result from the parser that created the node.
        Error messages are not included in the result. Use `self.content()`
        if the result plus any error messages is needed.
        """
418
419
        return self._result

420

421
    @result.setter
Eckhart Arnold's avatar
Eckhart Arnold committed
422
    def result(self, result: ResultType):
Eckhart Arnold's avatar
Eckhart Arnold committed
423
        # # made obsolete by static type checking with mypy
424
425
426
        # assert ((isinstance(result, tuple) and all(isinstance(child, Node) for child in result))
        #         or isinstance(result, Node)
        #         or isinstance(result, str)), str(result)
Eckhart Arnold's avatar
Eckhart Arnold committed
427
428
        # Possible optimization: Do not allow single nodes as argument:
        # assert not isinstance(result, Node)
429
430
        self._len = -1        # lazy evaluation
        self._content = None
431
432
433
434
435
436
437
438
439
        if isinstance(result, Node):
            self.children = (result,)
            self._result = self.children
        else:
            if isinstance(result, tuple):
                self.children = result
                self._result = result or ''
            else:
                self.children = NoChildren
eckhart's avatar
eckhart committed
440
                self._result = result  # cast(StrictResultType, result)
441

442
443

    @property
eckhart's avatar
eckhart committed
444
    def content(self) -> str:
445
        """
di68kap's avatar
di68kap committed
446
447
448
        Returns content as string, omitting error messages. If the node has
        child-nodes, the string content of the child-nodes is recursively read
        and then concatenated.
449
        """
450
451
452
453
        if self._content is None:
            if self.children:
                self._content = "".join(child.content for child in self.children)
            else:
454
                # self._content = self._result
455
456
457
                self._content = str(self._result)
                self._result = self._content  # self._result might be more efficient as a string!?
        return self._content
458
459


di68kap's avatar
di68kap committed
460
461
462
463
464
    @content.setter
    def content(self, content: str):
        self.result = content


465
466
467
468
469
470
471
472
473
    @property
    def structure(self) -> str:
        """
        Return structure (and content) as S-expression on a single line
        without any line breaks.
        """
        return flatten_sxpr(self.as_sxpr(showerrors=False))


474
    @property
475
    def pos(self) -> int:
eckhart's avatar
eckhart committed
476
477
        """Returns the position of the Node's content in the source text."""
        if self._pos < 0:
478
            raise AssertionError("Position value not initialized!")
479
480
        return self._pos

eckhart's avatar
eckhart committed
481

eckhart's avatar
eckhart committed
482
    def init_pos(self, pos: int) -> 'Node':
eckhart's avatar
eckhart committed
483
484
485
486
487
488
489
490
491
        """
        (Re-)initialize position value. Usually, the parser guard
        (`parsers.add_parser_guard()`) takes care of assigning the
        position in the document to newly created nodes. However,
        where Nodes are created outside the reach of the parser
        guard, their document-position must be assigned manually.
        This function recursively reassigns the position values
        of the child nodes, too.
        """
eckhart's avatar
eckhart committed
492
493
        assert self._pos < 0 or self.pos == pos, str("pos mismatch %i != %i" % (self._pos, pos))
        self._pos = pos
494
        # recursively adjust pos-values of all children
eckhart's avatar
eckhart committed
495
        offset = self.pos
496
        for child in self.children:
eckhart's avatar
eckhart committed
497
            child.init_pos(offset)
eckhart's avatar
eckhart committed
498
499
500
            offset = child.pos + len(child)
        return self

501

502
    @property
503
    def attr(self):
504
        """
505
        Returns a dictionary of XML-attr attached to the node.
506
        """
507
508
509
510
        if not hasattr(self, '_xml_attr'):
            self._xml_attr = OrderedDict()
        return self._xml_attr

Eckhart Arnold's avatar
Eckhart Arnold committed
511

512
    def _tree_repr(self, tab, open_fn, close_fn, data_fn=lambda i: i,
513
                   density=0, inline=False, inline_fn=lambda node: False) -> str:
514
515
516
517
518
519
520
521
        """
        Generates a tree representation of this node and its children
        in string from.

        The kind ot tree-representation that is determined by several
        function parameters. This could be an XML-representation or a
        lisp-like S-expression.

522
        Args:
523
            tab (str):  The indentation string, e.g. '\t' or '    '
eckhart's avatar
eckhart committed
524
            open_fn:   (Node->str) A function that returns an opening
525
                string (e.g. an XML-tag_name) for a given node
eckhart's avatar
eckhart committed
526
            close_fn:  (Node->str) A function that returns a closeF
527
                string (e.g. an XML-tag_name) for a given node.
eckhart's avatar
eckhart committed
528
            data_fn:   (str->str) A function that filters the data string
529
530
531
532
533
534
                before printing, e.g. to add quotation marks

        Returns (str):
            A string that contains a (serialized) tree representation
            of the node and its children.
        """
eckhart's avatar
eckhart committed
535
536
        head = open_fn(self)
        tail = close_fn(self)
537
538

        if not self.result:
539
            return head.rstrip() + tail.lstrip()
540

eckhart's avatar
eckhart committed
541
        tail = tail.lstrip(None if density & 2 else '')
542

543
        inline = inline or inline_fn(self)
544
545
546
547
548
        if inline:
            head = head.rstrip()
            tail = tail.lstrip()
            usetab, sep = '', ''
        else:
eckhart's avatar
eckhart committed
549
550
            usetab = tab if head else ''    # no indentation if tag is already omitted
            sep = '\n'
551

552
553
        if self.children:
            content = []
554
            for child in self.children:
555
                subtree = child._tree_repr(tab, open_fn, close_fn, data_fn,
556
                                           density, inline, inline_fn)
eckhart's avatar
eckhart committed
557
                if subtree:
eckhart's avatar
eckhart committed
558
559
                    st = [subtree] if inline else subtree.split('\n')
                    content.append((sep + usetab).join(s for s in st))
560
            return head + usetab + (sep + usetab).join(content) + tail
561

eckhart's avatar
eckhart committed
562
        res = self.content
eckhart's avatar
eckhart committed
563
        if not inline and not head:
eckhart's avatar
eckhart committed
564
            # strip whitespace for omitted non inline node, e.g. CharData in mixed elements
eckhart's avatar
eckhart committed
565
            res = res.strip()
566
567
        if density & 1 and res.find('\n') < 0:  # and head[0] == "<":
            # except for XML, add a gap between opening statement and content
568
            gap = ' ' if not inline and head and head.rstrip()[-1:] != '>' else ''
eckhart's avatar
eckhart committed
569
            return head.rstrip() + gap + data_fn(res) + tail.lstrip()
570
        else:
571
            return head + '\n'.join([usetab + data_fn(s) for s in res.split('\n')]) + tail
572

Eckhart Arnold's avatar
Eckhart Arnold committed
573

574
575
576
    def as_sxpr(self, src: str = None,
                showerrors: bool = True,
                indentation: int = 2,
577
                compact: bool = False) -> str:
578
579
580
        """
        Returns content as S-expression, i.e. in lisp-like form.

581
        Args:
582
583
584
            src:  The source text or `None`. In case the source text is
                given the position of the element in the text will be
                reported as line and column.
585
586
587
            showerrors: If True, error messages will be shown.
            indentation: The number of whitespaces for indentation
            compact:  If True, a compact representation is returned where
588
                brackets are omitted and only the indentation indicates the
Eckhart Arnold's avatar
Eckhart Arnold committed
589
                tree structure.
590
591
        """

eckhart's avatar
eckhart committed
592
        left_bracket, right_bracket, density = ('', '', 1) if compact else ('(', '\n)', 0)
eckhart's avatar
eckhart committed
593
        lbreaks = linebreaks(src) if src else []  # type: List[int]
594

595
        def opening(node) -> str:
eckhart's avatar
eckhart committed
596
            """Returns the opening string for the representation of `node`."""
eckhart's avatar
eckhart committed
597
            txt = [left_bracket, node.tag_name]
598
            # s += " '(pos %i)" % node.add_pos
599
            if hasattr(node, '_xml_attr'):
600
                txt.extend(' `(%s "%s")' % (k, v) for k, v in node.attr.items())
601
            if src:
602
603
                line, col = line_col(lbreaks, node.pos)
                txt.append(" `(pos %i %i %i)" % (node.pos, line, col))
604
            if showerrors and node.errors:
605
                txt.append(" `(err `%s)" % ' '.join(str(err) for err in node.errors))
606
            return "".join(txt) + '\n'
607

eckhart's avatar
eckhart committed
608
609
610
        def closing(node) -> str:
            """Returns the closing string for the representation of `node`."""
            return right_bracket
611

eckhart's avatar
eckhart committed
612
613
614
615
616
        def pretty(strg):
            """Encloses `strg` with the right kind of quotation marks."""
            return '"%s"' % strg if strg.find('"') < 0 \
                else "'%s'" % strg if strg.find("'") < 0 \
                else '"%s"' % strg.replace('"', r'\"')
617

618
        return self._tree_repr(' ' * indentation, opening, closing, pretty, density=density)
Eckhart Arnold's avatar
Eckhart Arnold committed
619

eckhart's avatar
eckhart committed
620

621
622
623
    def as_xml(self, src: str = None,
               showerrors: bool = True,
               indentation: int = 2,
eckhart's avatar
eckhart committed
624
625
626
               inline_tags: Set[str] = set(),
               omit_tags: Set[str] = set(),
               empty_tags: Set[str] = set()) -> str:
627
628
629
        """
        Returns content as XML-tree.

630
        Args:
631
632
633
            src:  The source text or `None`. In case the source text is
                given the position will also be reported as line and
                column.
634
635
            showerrors: If True, error messages will be shown.
            indentation: The number of whitespaces for indentation
636
637
638
            inline_tags:  A set of tag names, the content of which will always be written
                on a single line, unless it contains explicit line feeds ('\n').
            omit_tags:  A set of tags from which only the content will be printed, but
639
                neither the opening tag nor its attr nor the closing tag. This
640
641
642
                allows producing a mix of plain text and child tags in the output,
                which otherwise is not supported by the Node object, because it
                requires its content to be either a tuple of children or string content.
643
644
            empty_tags:  A set of tags which shall be rendered as empty elements, e.g.
                "<empty/>" instead of "<empty><empty>".
645
646
        """

647
        def opening(node) -> str:
648
649
650
            """Returns the opening string for the representation of `node`."""
            if node.tag_name in omit_tags:
                return ''
eckhart's avatar
eckhart committed
651
            txt = ['<', node.tag_name]
652
            has_reserved_attrs = hasattr(node, '_xml_attr') \
eckhart's avatar
eckhart committed
653
                and any(r in node.attr for r in {'err', 'line', 'col'})
di68kap's avatar
di68kap committed
654
            if hasattr(node, '_xml_attr'):
655
                txt.extend(' %s="%s"' % (k, v) for k, v in node.attr.items())
656
            if src and not has_reserved_attrs:
eckhart's avatar
eckhart committed
657
                txt.append(' line="%i" col="%i"' % line_col(line_breaks, node.pos))
658
            if showerrors and node.errors and not has_reserved_attrs:
eckhart's avatar
eckhart committed
659
660
                txt.append(' err="%s"' % ''.join(str(err).replace('"', r'\"')
                                                 for err in node.errors))
661
662
663
            if node.tag_name in empty_tags:
                assert not node.result, ("Node %s with content %s is not an empty element!" %
                                         (node.tag_name, str(node)))
eckhart's avatar
eckhart committed
664
                ending = "/>\n" if not node.tag_name[0] == '?' else "?>\n"
665
666
667
            else:
                ending = ">\n"
            return "".join(txt + [ending])
668
669

        def closing(node):
670
            """Returns the closing string for the representation of `node`."""
671
            if node.tag_name in omit_tags or node.tag_name in empty_tags:
672
                return ''
673
            return ('\n</') + node.tag_name + '>'
674

675
676
677
678
679
680
681
        def sanitizer(content: str) -> str:
            """Substitute "&", "<", ">" in XML-content by the respective entities."""
            content = RX_AMP.sub('&amp;', content)
            content = content.replace('<', '&lt;').replace('>', '&gt;')
            return content


682
683
684
685
686
        def inlining(node):
            """Returns True, if `node`'s tag name is contained in `inline_tags`,
            thereby signalling that the children of this node shall not be
            printed on several lines to avoid unwanted gaps in the output.
            """
eckhart's avatar
eckhart committed
687
688
689
            return node.tag_name in inline_tags \
                or (hasattr(node, '_xml_attr')
                    and node.attr.get('xml:space', 'default') == 'preserve')
690

691
        line_breaks = linebreaks(src) if src else []
692
        return self._tree_repr(' ' * indentation, opening, closing, sanitizer,
693
                               density=1, inline_fn=inlining)
694

Eckhart Arnold's avatar
Eckhart Arnold committed
695

eckhart's avatar
eckhart committed
696
    def select(self, match_function: Callable, include_root: bool = False, reverse: bool = False) \
697
            -> Iterator['Node']:
eckhart's avatar
eckhart committed
698
        """
699
        Finds nodes in the tree that fulfill a given criterion.
eckhart's avatar
eckhart committed
700

701
        `select` is a generator that yields all nodes for which the
di68kap's avatar
di68kap committed
702
        given `match_function` evaluates to True. The tree is
703
704
705
        traversed pre-order.

        See function `Node.select_by_tag` for some examples.
eckhart's avatar
eckhart committed
706

707
708
709
        Args:
            match_function (function): A function  that takes as Node
                object as argument and returns True or False
710
711
            include_root (bool): If False, only descendant nodes will be
                checked for a match.
712
713
            reverse (bool): If True, the tree will be walked in reverse
                order, i.e. last children first.
714
        Yields:
715
            Node: All nodes of the tree for which
716
717
            ``match_function(node)`` returns True
        """
718
        if include_root and match_function(self):
719
            yield self
720
721
722
        child_iterator = reversed(self.children) if reverse else self.children
        for child in child_iterator:
            for node in child.select(match_function, True, reverse):
eckhart's avatar
eckhart committed
723
                yield node
724

Eckhart Arnold's avatar
Eckhart Arnold committed
725

eckhart's avatar
eckhart committed
726
    def select_by_tag(self, tag_names: Union[str, AbstractSet[str]],
eckhart's avatar
eckhart committed
727
                      include_root: bool = False) -> Iterator['Node']:
728
        """
729
730
731
732
        Returns an iterator that runs through all descendants that have one
        of the given tag names.

        Examples::
733

734
            >>> tree = parse_sxpr('(a (b "X") (X (c "d")) (e (X "F")))')
735
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag("X", False))
736
            ['(X (c "d"))', '(X "F")']
737
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag({"X", "b"}, False))
738
            ['(b "X")', '(X (c "d"))', '(X "F")']
739
            >>> any(tree.select_by_tag('a', False))
740
            False
741
            >>> list(flatten_sxpr(item.as_sxpr()) for item in tree.select_by_tag('a', True))
742
            ['(a (b "X") (X (c "d")) (e (X "F")))']
743
744
            >>> flatten_sxpr(next(tree.select_by_tag("X", False)).as_sxpr())
            '(X (c "d"))'
745
746

        Args:
747
748
749
750
            tag_name(set): A tag name or set of tag names that is being
                searched for
            include_root (bool): If False, only descendant nodes will be
                checked for a match.
751
752
753
        Yields:
            Node: All nodes which have a given tag name.
        """
754
        if isinstance(tag_names, str):
eckhart's avatar
eckhart committed
755
            tag_names = frozenset({tag_names})
756
        return self.select(lambda node: node.tag_name in tag_names, include_root)
757
758


759
760
761
762
    def pick(self, tag_names: Union[str, Set[str]]) -> Optional['Node']:
        """
        Picks the first descendant with one of the given tag_names.

763
        This function is mostly just syntactic sugar for
764
765
766
767
768
769
770
771
772
773
        ``next(node.select_by_tag(tag_names, False))``. However, rather than
        raising a StopIterationError if no descendant with the given tag-name
        exists, it returns None.
        """
        try:
            return next(self.select_by_tag(tag_names, False))
        except StopIteration:
            return None


774
    def tree_size(self) -> int:
eckhart's avatar
eckhart committed
775
776
777
        """
        Recursively counts the number of nodes in the tree including the root node.
        """
778
779
780
        return sum(child.tree_size() for child in self.children) + 1


di68kap's avatar
di68kap committed
781
782
783
ZOMBIE_NODE = Node(ZOMBIE_PARSER, '')


di68kap's avatar
di68kap committed
784
class RootNode(Node):
785
    """TODO: Add Documentation!!!
di68kap's avatar
di68kap committed
786

Eckhart Arnold's avatar
Eckhart Arnold committed
787
        all_errors (list):  A list of all errors that have occured so far during
788
789
                processing (i.e. parsing, AST-transformation, compiling)
                of this tree.
di68kap's avatar
di68kap committed
790

791
792
        error_flag (int):  the highest warning or error level of all errors
                that occurred.
di68kap's avatar
di68kap committed
793
    """
794

eckhart's avatar
eckhart committed
795
    def __init__(self, node: Optional[Node] = None):
di68kap's avatar
di68kap committed
796
        super().__init__(ZOMBIE_PARSER, '')
eckhart's avatar
eckhart committed
797
        self.all_errors = []  # type: List[Error]
di68kap's avatar
di68kap committed
798
        self.error_flag = 0
eckhart's avatar
eckhart committed
799
800
        if node is not None:
            self.swallow(node)
801
        # customization for XML-Representation
eckhart's avatar
eckhart committed
802
803
804
        self.inline_tags = set()  # type: Set[str]
        self.omit_tags = set()  # type: Set[str]
        self.empty_tags = set()  # type: Set[str]
di68kap's avatar
di68kap committed
805

806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
    def __deepcopy__(self, memodict={}):
        duplicate = self.__class__(None)
        if self.children:
            duplicate.children = copy.deepcopy(self.children)
            duplicate._result = duplicate.children
        else:
            duplicate.children = NoChildren
            duplicate._result = self._result
        duplicate.errors = copy.deepcopy(self.errors) if self.errors else []
        duplicate._pos = self._pos
        duplicate._len = self._len
        if hasattr(self, '_xml_attr'):
            duplicate._xml_attr = copy.deepcopy(self._xml_attr)
        duplicate.all_errors = copy.deepcopy(self.all_errors)
        duplicate.error_flag = self.error_flag
        duplicate.inline_tags = self.inline_tags
        duplicate.omit_tags = self.omit_tags
        duplicate.empty_tags = self.empty_tags
        duplicate.parser = self.parser
        return duplicate


828
    def swallow(self, node: Node) -> 'RootNode':
829
830
        """
        Put `self` in the place of `node` by copying all its data.
831
832
833
834
835
836
837
838
839
        Returns self.

        This is done by the parse.Grammar object after
        parsing has finished, so that the Grammar object always
        returns a syntax tree rooted in a RootNode object.

        It is possible to add errors to a RootNode object, before it
        has actually swallowed the root of the syntax tree.
        """
di68kap's avatar
di68kap committed
840
841
842
843
844
        self._result = node._result
        self.children = node.children
        self._len = node._len
        self._pos = node._pos
        self.parser = node.parser
845
846
        if hasattr(node, '_xml_attr'):
            self._xml_attr = node._xml_attr
di68kap's avatar
di68kap committed
847
        self._content = node._content
848
        return self
di68kap's avatar
di68kap committed
849

eckhart's avatar
eckhart committed
850
    def add_error(self, node: Node, error: Error) -> 'RootNode':
851
852
853
        """
        Adds an Error object to the tree, locating it at a specific node.
        """
di68kap's avatar
di68kap committed
854
855
856
857
858
        self.all_errors.append(error)
        self.error_flag = max(self.error_flag, error.code)
        node.errors.append(error)
        return self

eckhart's avatar
eckhart committed
859
    def new_error(self,
860
                  node: Node,
di68kap's avatar
di68kap committed
861
                  message: str,
eckhart's avatar
eckhart committed
862
                  code: ErrorCode = Error.ERROR) -> 'RootNode':
di68kap's avatar
di68kap committed
863
        """
eckhart's avatar
eckhart committed
864
        Adds an error to this tree, locating it at a specific node.
di68kap's avatar
di68kap committed
865
866
867
868
869
        Parameters:
            pos(int):     The position of the error in the source text
            message(str): A string with the error message.abs
            code(int):    An error code to identify the kind of error
        """
Eckhart Arnold's avatar
Eckhart Arnold committed
870
        error = Error(message, node.pos, code)
eckhart's avatar
eckhart committed
871
872
        self.add_error(node, error)
        return self
di68kap's avatar
di68kap committed
873

eckhart's avatar
eckhart committed
874
    def collect_errors(self) -> List[Error]:
875
876
        """
        Returns the list of errors, ordered bv their position.
di68kap's avatar
di68kap committed
877
        """
878
        self.all_errors.sort(key=lambda e: e.pos)
Eckhart Arnold's avatar
Eckhart Arnold committed
879
        return self.all_errors
di68kap's avatar
di68kap committed
880

881
    def customized_XML(self):
882
883
        """
        Returns a customized XML representation of the tree.
884
        See the docstring of `Node.as_xml()` for an explanation of the
885
886
        customizations.
        """
eckhart's avatar
eckhart committed
887
        return self.as_xml(inline_tags=self.inline_tags,
888
889
890
                           omit_tags=self.omit_tags,
                           empty_tags=self.empty_tags)

di68kap's avatar
di68kap committed
891

eckhart's avatar
eckhart committed
892
893
ZOMBIE_ROOTNODE = RootNode()

di68kap's avatar
di68kap committed
894
895
896
897
898
#######################################################################
#
# S-expression- and XML-parsers
#
#######################################################################
eckhart's avatar
eckhart committed
899
900


eckhart's avatar
eckhart committed
901
def parse_sxpr(sxpr: Union[str, StringView]) -> Node:
902
    """
903
904
905
906
907
    Generates a tree of nodes from an S-expression.

    This can - among other things - be used for deserialization of trees that
    have been serialized with `Node.as_sxpr()` or as a convenient way to
    generate test data.
908
909

    Example:
910
    >>> parse_sxpr("(a (b c))").as_sxpr()
911
    '(a\\n  (b\\n    "c"\\n  )\\n)'
912
    """
913

eckhart's avatar
eckhart committed
914
    sxpr = StringView(sxpr).strip() if isinstance(sxpr, str) else sxpr.strip()
eckhart's avatar
eckhart committed
915
    mock_parsers = dict()  # type: Dict[str, MockParser]